This section provides background information related to the present disclosure which is not necessarily prior art.
With the development of communication technology, software can implement various functions. Software performance directly impacts user experience. Thus, the software performance needs to be evaluated to improve the software performance.
There are various methods for determining software performance.
In first method, before the software is published the software is run on multiple computers in a laboratory environment to acquire resource occupancy conditions, such as CPU occupancy rate, the number of IO reading/writing, the size of memory, and so on when the software is running on each computer. The software performance is determined according to the CPU occupancy rate, the number of the 10 reading/writing, and the size of the memory when the software is running on each computer. In second method, after the software is published, time consumed on key steps of running the software is acquired during a process of running the software. According to the time assumed on the key steps, the software performance is determined.
There are the following shortcomings in the conventional technology. In the first method, since the resource occupancy conditions, such as the CPU occupancy rate, the number of the IO reading/writing, the size of the memory, and so on when the software is running on each computer is acquired in the laboratory environment and the number of the computers in the laboratory are limited, the software performance determined according to the first method is not accurate. In the second method, it is hard to configure the key steps of running the software. Furthermore, since the software performance is determined only according to the key steps rather than the whole process of running the software, the software performance determined according to the second method is not accurate.